Summary and Key References

This module traced the full path from a database of individual reports to a defensible regulatory decision: where signals originate, how disproportionality analysis surfaces statistical candidates at scale, how Bayesian methods keep small case counts from generating an unmanageable flood of false alarms, and how a qualified human reviewer turns a statistical flag into a validated signal by weighing biological plausibility and ruling out confounding explanations a formula can’t see.

From there, prioritisation, the EU’s PRAC process as a concrete worked example, a comparison across the three major global databases, and the genuine but bounded role of AI/ML methods all built toward the same closing idea: signal management is a loop, not a one-way pipeline. Its output — a label update, an additional risk minimisation measure, or a documented decision not to act — feeds directly back into how every future case gets assessed, which is exactly the connective tissue linking this module to Module 6’s expectedness lesson and forward to Module 9’s risk management content.

Key references: GVP Module IX (Signal Management), ICH E2E (Pharmacovigilance Planning), WHO-UMC/Uppsala Monitoring Centre disproportionality methodology documentation, EU Regulation 2025/1466, CIOMS Working Group XIV Final Report.